ElevenLabs, founded in 2022, launched its text-to-speech and voice-cloning platform in beta in January 2023 and came out of beta on August 22, 2023 with Eleven Multilingual v2, a foundational speech model supporting nearly 30 languages. The system could automatically detect the written language and generate emotionally expressive speech, and crucially it preserved a speaker’s unique voice characteristics - including their original accent - across all of those languages. The same release introduced Professional Voice Cloning, which the company said could create a digital copy of a person’s voice virtually indistinguishable from the original.
The quality leap is what made ElevenLabs a landmark. Earlier text-to-speech sounded robotic; ElevenLabs produced audio with natural prosody, pacing, and emotion good enough for audiobooks, dubbing, games, and video narration. That same realism made it the tool of choice for a flood of audio deepfakes - cloned celebrity and politician voices spread online within weeks of the public launch, and AI-voiced robocalls became a concrete fraud and election-integrity worry.
The technology helped drive a regulatory response. In 2024 the US FCC ruled AI-generated voices in robocalls illegal, and Tennessee passed the ELVIS Act to protect a person’s voice against unauthorized AI cloning. ElevenLabs itself added voice-verification and provenance safeguards in response to the misuse.
Why business readers should care: realistic voice cloning is simultaneously a major accessibility and localization tool and a potent fraud vector. The same capability that lets a company dub a video into 30 languages also lets a scammer impersonate an executive on the phone, which is why voice is now a regulated frontier of synthetic media.